Bridging Commonsense Reasoning and Probabilistic Planning via a Probabilistic Action Language
This work addresses the problem of reducing engineering complexity for AI researchers and practitioners in sequential decision-making tasks, though it is incremental as it builds upon the existing icorpp framework.
The paper tackles the challenge of integrating commonsense reasoning with probabilistic planning by extending the probabilistic action language pBC+ to unify these components, resulting in a system that automatically compiles action descriptions into POMDP models for optimal policy computation without manual engineering efforts.
To be responsive to dynamically changing real-world environments, an intelligent agent needs to perform complex sequential decision-making tasks that are often guided by commonsense knowledge. The previous work on this line of research led to the framework called "interleaved commonsense reasoning and probabilistic planning" (icorpp), which used P-log for representing commmonsense knowledge and Markov Decision Processes (MDPs) or Partially Observable MDPs (POMDPs) for planning under uncertainty. A main limitation of icorpp is that its implementation requires non-trivial engineering efforts to bridge the commonsense reasoning and probabilistic planning formalisms. In this paper, we present a unified framework to integrate icorpp's reasoning and planning components. In particular, we extend probabilistic action language pBC+ to express utility, belief states, and observation as in POMDP models. Inheriting the advantages of action languages, the new action language provides an elaboration tolerant representation of POMDP that reflects commonsense knowledge. The idea led to the design of the system pbcplus2pomdp, which compiles a pBC+ action description into a POMDP model that can be directly processed by off-the-shelf POMDP solvers to compute an optimal policy of the pBC+ action description. Our experiments show that it retains the advantages of icorpp while avoiding the manual efforts in bridging the commonsense reasoner and the probabilistic planner.